Research on the classifier with the tree frame based on multiple attractor cellular automaton

The partition of a pattern space as the view of a cell space is a uniform partition, it is difficult to adapt to the needs of spatial non-uniform partition. In this paper, a cellular automaton classifier with a tree structure is constructed by combing with the CART algorithm. The construction method of the characteristic matrix of the multiple attractor cellular automata is studied based on the particle swarm optimization method, and this method can build the nodes of the multiple attractor cellular automata. This kind of classifier can solve the non-uniform partition problem and obtain a good classification performance while using a pseudo-exhaustive field with less bits. The experiment results show that our algorithm is more accurate than those obtained through the multiple attractor cellular automata.